Liquid crystal (LC) droplets dispersed in aqueous solution are an interesting stimuli-responsive material [1] . They have been used as a model system for studying particle rotations [2] [3] , optical momentum transfers [4] , and optical vortex generations [5] . In recent years, LC droplets have also emerged as a unique optical probe for the detection of chemical and biological species and their reactions at the surface of the droplets [6] . The director configuration of LC droplets is known to reflect the balance between the elasticity and the surface anchoring of the LC inside the droplets. Thus, the adsorption of chemical and biological species at the LC/aqueous interface may disrupt the balance to trigger the configuration transition of the LC inside the droplets, which can be easily observed with a polarizing optical microscope. The ultra-sensitivity of naked LC droplets has been proven in sensing bacterial endotoxin in aqueous solution [7] . However, naked LC droplets in aqueous solution are unstable and tend to coalesce over time. There has been interest in stabilizing and functionalizing LC droplets in aqueous solution by the adsorption of polymers at the LC/aqueous interface for sensor applications [8] - [20] . It has been shown that the adsorption of polyelectrolytes at the LC/aqueous interface can stabilize the LC droplets in aqueous solution [10] [13] [14] . Polyelectrolyte-coated LC droplets can be further modified by the adsorption of oppositely charged polyelectrolytes. It has been shown that the multilayer coating formed by the layer-by- layer adsorption of negatively charged poly(styrenesulfonate sodium) (PSS) and positively charged poly(allylamine hydrochloride) (PAH) is permeable for surfactants [10] . The permeated surfactants are able to react with the LC droplet cores and induce the bipolar-to-radial configuration transition of the LC inside the droplet cores. However, the PSS/PAH multilayer coating is found to reduce the response time of the LC droplet cores to surfactants by two orders of magnitude compared with naked LC droplets dispersed in aqueous solution.

In this paper, we coat 4-n-pentyl-4’-cyanobiphenyl (5CB) droplets with a bipolar configuration by the adsorption of both positively and negatively charged polyelectrolytes at the 5CB/aqueous interface to study the influence of polyelectrolyte coatings on the interaction of surfactants with the 5CB droplet cores by observing their configuration transitions as a function of surfactant concentrations. Polyelectrolytes used in our experiments are positively charged chitosan (CHT) and negatively charged polystyrene sulfonate (PSS). Surfactants used in our experiments are anionic sodium alkyl sulfate (SC_nS, n = 18, 16, 14, and 12), cationic alkyl trimethylammonium bromide (C_nTAB, n = 18, 16, 14, and 12) and nonionic tetraethylene glycol alkyl ether (C_nE₄, n = 14 and 12). Although these surfactants are able to penetrate into the polyelectrolyte coatings to induce the bipolar-to-radial transition of the LC droplet cores, the concentration of charged surfactants required to induce the bipolar-to- radial configuration transition of the 5CB droplet cores is affected by charged polyelectrolyte coating. The effect of charged polyelectrolyte coatings is significantly enlarged with the decrease of the alkyl chain length of charged surfactants.

The chemical structures of CHT, PSS, SC_nS, C_nTAB, and C_nE₄ are shown in Figure 1. 5CB droplets were formed in deionized water with positively charged CHT or negatively charged PSS at pH 5.4. The adsorption of these charged polyelectrolytes at the 5CB/water interface is evident from ζ-potential measurements. The ζ-potentials of CHT and PSS-coated 5CB droplets are +38 mV and –43 mV, respectively. The average diameter of CHT and PSS-coated 5CB droplets is ~0.70 ± 0.15 μm. Figure 2(a) shows a polarizing optical microscopy image of positively charged CHT-coated 5CB droplets in water. The CHT-coated 5CB droplets are stable and show a bipolar configuration, suggesting a parallel surface anchoring of the 5CB in the droplet cores. After being incubated with negatively charged SC₁₂S in water for 30 min, the CHT-coated 5CB droplets gradually transit into a radial

configuration (Figure 2(b) and Figure 2(c)), suggesting a perpendicular anchoring of the 5CB in the droplet cores. In addition, we note that the ζ-potentials of CHT-coated 5CB droplets drop from +38 mV to +12 mV after 30 min incubation with negatively charged SC₁₂S. It has been shown that the adsorption of SC₁₂S at the surface of naked 5CB droplets dispersed in water can induce the bipolar-to-radial configuration transition of naked 5CB droplets [21] [22] . Thus, we can conclude that negatively charged SC₁₂S penetrates into the positively charged CHT coating and react with the 5CB droplet cores, inducing the bipolar-to-radial transition shown in Figure 2. It is also visible from Figure 2(b) that the bipolar-to-radial transition of small CHT-coated 5CB droplets is faster than large CHT-coated 5CB droplets.

Figure 3(a) shows the transition curves of positively charged CHT-coated 5CB droplets as a function of SC_nS concentrations, in which the total number of the droplets is kept the same (~8.2 × 10⁸ mL⁻¹). For n = 12, the bipolar-to-radial configuration transition of CHT-coated 5CB droplets starts at the SC₁₂S concentration of ~14 µM. all the CHT-coated 5CB droplets transit into the radial configuration when the concentration of SC₁₂S reaches to ~28 µM. The transition curves of CHT-coated 5CB droplets gradually shift to left with the increase of n (Figure 3(a)). For n = 18, the 100% transition of CHT-coated 5CB droplets occurs at the SC₁₈S concentration of ~4 µM, which is seven times lower than the SC₁₂S concentration (~28 µM) required to induce the 100% transition of CHT-coated 5CB droplets. These concentrations are significantly lower than the critical micelle concentration of SC₁₂S (7 mM) and SC₁₈S (0.1 mM), suggesting that no aggregation occurs for SC₁₂S and SC₁₈S, which were used in our experiments. The chain length-dependent transition should be related to the anchoring energy of SC_nS, which increases with the increase of its alkyl chain lengths [23] . Thus, we can expect that CHT- coated 5CB droplets are more susceptible to SC_nS with longer alkyl chain lengths, which explain why the concentration of SC_nS required inducing the bipolar-to-radial transition of CHT-coated 5CB droplets decreases with the increase of SC_nS chain lengths. Figures 3(b) shows that the transition curves of negatively charged PSS- coated 5CB droplets as a function of SC_nS concentrations, in which the total number of the droplets remains the same (~8.2 × 10⁸ mL⁻¹). The transition curves gradually shift to left with the increase of n, which is similar to the positively charged CHT-coated droplets. However, the concentration of negatively charged SC_nS required to induce the bipolar-to-radial transition of negatively charged PSS-coated 5CB droplets is larger than that of negatively charged CHT-coated 5CB droplets. This is a result of the reduced penetration of negatively charged SC_nS into the negatively charged PSS-coated 5CB droplets due to electrostatic repulsion because the anchoring energy depends on the density of the SC_nS adsorbed at the surface of 5CB droplets. It has been shown that the density increase of surfactant layers favors the planar-to-perpendicular anchoring transition of LCs [24] - [26] . For the comparison, we plot the SC_nS concentration required to induce the 100% transition of CHT- and PSS-coated 5CB droplets as a function of n (Figure 3(c)). It is clear from Figure 3(c) that the difference is enlarged with the decrease of n. The concentration of SC₁₈S required to induce the 100% transition of CHT- and PSS-coated 5CB droplets is 4 µM and 12 µM, respectively. While the concentration of SC₁₂S required inducing the 100% transition of CHT- and PSS-coated 5CB droplets is 28 µM and 200 µM, respectively. This result suggests that the polyelectrolyte coating has more significant impact for the interaction between the 5CB droplet cores and the

SC_nS with shorter alkyl chain lengths. The negatively charged head group of SC_nS is expected to locate at the outside of the negatively charged PSS coating due to electrostatic repulsion. The extent of the penetrated alkyl chain of SC_nS into the 5CB droplet cores decreases with the decrease of their lengths. Thus, the effect of the PSS coating increases with the decrease of the alkyl chain length of SC_nS.

It has shown that the adsorption of alkyl trimethylammonium bromide (C_nTAB) with n = 18, 16, 14, and 12 at the LC/aqueous interface can also induce the bipolar-to-radial director configuration transition of naked LC droplets [22] . The chain length-dependent transition is also observed for CHI- and PSS-coated 5CB droplets after being exposed to positively charged C_nTAB. Figure 4(a) shows the transition curves of CHT-coated 5CB droplets from the bipolar to the radial configuration as a function of C_nTAB concentrations, in which the total number of the droplets is kept the same (~8.2 × 10⁸ mL⁻¹). For n = 12, the configuration transition of CHT- coated 5CB droplets starts at the C₁₂TAB concentration of 200 µM. All the CHT-coated 5CB droplets transit to the radial configuration when the concentration of C₁₂TAB reaches to 270 µM. The transition curves shift to left with the increase of n, The concentration of C_nTAB required to induce 100% of the transition drops from ~270 µM to ~75 µM when the chain length of C_nTAB increases from n = 12 to n = 18. These concentrations are significantly lower than the critical micelle concentration of C₁₂TAB (15mM) and C₁₈TAB (0.9 mM), suggesting that there is no aggregation for SC₁₂S and SC₁₈S, which were used in our experiments. Figure 4(b) shows the transition curves of negatively charged PSS-modified 5CB droplets as a function of C_nTAB concentrations. The adsorption of C₁₂TAB on the PSS-modified 5CB droplets was confirmed by the decrease of the ζ-potentials of CHT-coated 5CB droplets from −43 mV to −11 mV after 30 min incubation with the positive charged C₁₂TAB. The transition curves are shifted leftwards with the increase of n. It is clear from Figure 4(c) that the concentration of C_nTAB required to induce the 100% transition of PSS-coated 5CB droplets is much lower than that of the CHT-coated 5CB droplets. For n = 18, the 100% transition of PSS-coated 5CB droplets takes place at the C₁₈TAB concentration of ~10 µM, while is lower than the concentration of C₁₈TAB required to induce the 100% transition of CHT-coated 5CB droplets (~75 µM). Again, the differences are significantly enlarged with the decrease of n.

Furthermore, we study the interaction of nonionic C_nE₄ with positively charged CHT and negatively charged PSS-coated 5CB droplets. For n = 12 or 14, there are only slightly shifts observed in the transition curves between positively charged CHT-coated 5CB droplets and negatively charged PSS-coated 5CB droplets (Figure 5(a)). The concentration of SC₁₄S required to induce the 100% transition of CHT- and PSS-coated 5CB droplets is near the same. While the concentration of SC₁₂S required inducing the 100% transition of PSS-coated 5CB droplets is slightly larger than that of CHT-coated 5CB droplets (Figure 5(b)). This result suggests that the charged polyelectrolyte-coating has no significantly effect on the interaction between the 5CB droplet cores and nonionic C_nE₄. Finally, we compare the transition curves of positively charged CHT-coated 5CB droplets for anionic C₁₂TAB, ionic SC₁₂S, and nonionic C₁₂E₄ (Figure 6(a)). The concentration of nonionic C₁₂E₄ required to induce the 100% transition of PSS-coated 5CB droplets is near the same as cationic C₁₂TAB, but significantly lower than anionic SC₁₂S. Similar results are observed for anionic C₁₄TAB, ionic SC₁₄S, and nonionic C₁₄E₄ (Figure 6(b)).

We have shown that the surfactant-induced configuration transition of charged polyelectrolyte-coated 5CB drop- lets is affected by the electrostatic interaction between charged surfactants and charged polyelectrolyte coatings. The charged polyelectrolyte coating causes the increase of the concentration of surfactants with the same charge, which is required to induce the bipolar-to-radial transition of the LC droplet cores. The effect of charged polyelectrolyte coatings is significantly enlarged with the decrease of surfactant chain lengths. Our results suggest the possibility of engineering polyelectrolyte coatings to tune the interaction of LC droplet cores with surfactants, which is critical towards designing LC droplet based sensors.

This work was supported by USA National Science Foundation (CBET-1264355).